JP2001251627A - Encoding device, encoding method, and recording medium recording program - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide an encoding device with improved encoding efficiency of moving image data. Kind Code: A1 A coding apparatus for coding a prediction error and a motion vector obtained by performing a motion compensation prediction by a motion prediction unit and a motion vector detection unit on a block basis of moving image data. Part 1
20 determines whether or not the DCT and the prediction error of the block to be coded after quantization are 0, and when it is determined that the prediction error is 0, the motion vector of the P picture is smaller than the first threshold value. If the B-picture is smaller, the motion vector replacement unit 121 is notified to omit the block if the difference between the size of the B-picture and the motion vector of the adjacent block is smaller than the second threshold, and the motion vector replacement unit 121 By replacing the motion vector from the motion vector detection unit 102 with 0 and outputting it to the variable length coding unit 110, the variable length coding unit 110 omits variable length coding of the block.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding apparatus for compressing and encoding moving image data, and more particularly to an improvement in encoding efficiency.

[0002]

2. Description of the Related Art An encoding apparatus for compressing and encoding moving image data by the MPEG system compresses moving image data by performing inter-frame prediction by motion compensation (hereinafter abbreviated as motion compensation prediction). Motion compensation prediction refers to a block to be encoded in video data (hereinafter, referred to as a target block).
, A block (hereinafter, referred to as a reference block) similar to the target block is searched for from among the past and / or future (or both) frames of the frame to which the target block belongs (hereinafter, referred to as a reference frame). It refers to an operation of obtaining a difference between each pixel between the two blocks (hereinafter, referred to as a prediction error) and a motion vector representing a relative positional relationship between the two blocks. Normally, motion compensation prediction is performed for each macroblock. The macro block is composed of four blocks of luminance signals of 8 pixels × 8 lines and two blocks of color difference signals of 8 pixels × 8 lines at positions corresponding to the four blocks.

An encoding device performs DCT (Discrete Cosine Transform) and quantization on a prediction error obtained by motion compensation prediction, and obtains a quantized DCT coefficient block (hereinafter, referred to as a DCT coefficient block).
(Referred to as a quantization block) together with a motion vector. However, the coding apparatus does not perform motion compensation prediction on the macroblock of the I picture among the three types of coding types (I, P, and B pictures). ,
Perform variable length coding.

Here, the coding type will be described. There are three types of coding types: I picture, P picture, and B picture. For each frame, one of the three coding types is determined, and different coding is performed according to the coding type. Encoded by the method.
In the frame of the I picture, the luminance signal of the macroblock and the two chrominance signals themselves are D signals without performing motion compensation prediction.
CT, quantization and variable length coding. The I-picture frame is a reference frame for a frame of another encoding type. A P-picture frame is motion-compensated and predicted using a past I-picture or P-picture frame as a reference frame, and the prediction error after the motion-compensated prediction is subjected to DCT, quantization, and variable-length coding. The P picture frame is a reference frame for another P picture or B picture macroblock. The encoding method for P pictures is called unidirectional prediction. A B-picture frame is motion-compensated and predicted using a past or future (or both) I-picture or P-picture frame as a reference frame, and the resulting prediction error is subjected to DCT, quantization, and variable-length coding. It does not become a reference frame for other frames. The encoding method of a B picture is called bidirectional prediction because it is predicted from both past and future directions.

[0005] In addition to the above series of operations, the encoding apparatus calls a macroblock skip, and when a quantized block or a motion vector has the same value as an adjacent macroblock or a macroblock at the same position in the previous and next frames, the variable length is used. An operation to omit encoding is being performed. Here, a macroblock in which a macroblock is skipped is referred to as a skipped macroblock.

[0006] More specifically, in the case of a P-picture, the coding apparatus uses a macroblock for a macroblock in which the quantized block is 0, that is, all the quantized DCT coefficients in the quantized block are 0 and the motion vector is 0. Block skip is performed, and in the case of a B picture, the quantization block is 0
The macroblock skip is performed for a macroblock whose motion vector is the same in magnitude and direction as the motion vector of the macroblock adjacent to the left of the macroblock. However, the encoding device does not skip the macroblocks located at the beginning and end of the slice.

[0007] Here, a slice is one layer of a hierarchical structure in the encoding of moving image data, and is composed of ten or more macroblocks arranged side by side on a frame. This is stipulated by the standard to realize random access in a decoding device that decodes moving image data, and the decoding device skips a slice and decodes the beginning of the next slice even if part of the moving image data is broken be able to. Since the head and the end of the slice are required for the random access, the encoding apparatus does not skip those macroblocks.

[0008] When the decoding apparatus for decoding moving picture data detects that there is a skipped macroblock from the address of the macroblock, the decoding apparatus decodes the P picture using the block located at the same position as the skipped macroblock in the reference frame of the picture. However, the B picture is decoded using the motion vector of the macroblock on the left.

[0009]

In a conventional coding apparatus, even if a macroblock is skipped, there is a high probability that a macroblock having a small degree of deterioration will be coded without skipping the macroblock. That is, a macroblock with a low degree of deterioration is not always a macroblock to be skipped, so that there is a problem that the compression efficiency is low.

Accordingly, an object of the present invention is to provide an encoding apparatus, an encoding method, and a recording medium on which an encoding process is recorded, in which the compression rate is improved and the amount of encoding processing is reduced.

[0011]

An encoding apparatus according to the present invention comprises:
An encoding apparatus that performs encoding including motion compensation prediction on an encoding target block for each block of moving image data, and illustrates how conspicuous pixel shift occurs when encoding of the encoding target block is omitted. The control device includes a calculating unit that calculates an index, and a control unit that controls to omit encoding of the current block according to the index.

According to this configuration, the encoding apparatus omits the encoding of the block to be encoded, that is, encodes the block in which the pixel shift is conspicuous, according to the index indicating the degree of conspicuousness of the pixel shift, Is omitted because the encoding of the block that is not noticeable is omitted, so that even if the pixel deviation due to the encoding omission is small as in the conventional encoding device, the block is encoded without being omitted. This eliminates inconsistency with the omission of encoding, and has the effect of reducing the amount of encoding processing and increasing the compression ratio.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An encoding apparatus according to an embodiment of the present invention will be described below with reference to the drawings. <First Embodiment> FIG. 1 shows an encoding apparatus 10 according to a first embodiment.
FIG. 3 is a block diagram showing the configuration of FIG. In FIG. 1, an encoding device 10 includes an input terminal 100, an image memory unit 101, a motion vector detection unit 102, a difference unit 103, a DCT unit 104,
Quantization unit 105, inverse quantization unit 106, inverse DCT unit 10
7, an adder 108, a motion predictor 109, a variable length encoder 110, a buffer 111, an output terminal 112, a skip determiner 120, and a motion vector replacer 121,
The moving image data input from the outside via the input terminal 100 is compression-encoded and a bit string is output from the output terminal 112.

The image memory unit 101 holds moving image data input from the outside via the input terminal 100. The moving image data is output from the image memory unit 101 to the difference unit 103 on a macroblock basis, with the order of the frames rearranged in the encoding order within the image memory unit 101.
The control unit 113 rearranges the frames in coding order according to the picture type.

The motion vector detection unit 102 detects a motion vector of the moving image data stored in the image memory unit 101 by motion compensation prediction on a macroblock basis.
Output to the motion prediction unit 109, the skip determination unit 120, and the motion vector replacement unit 121. The part including the difference unit 103, the DCT unit 104, the quantization unit 105, the skip determination unit 120, and the variable length coding unit 110 determines whether to calculate prediction error, DCT, quantize, and skip macroblocks in macroblock units. Is determined, and variable-length coding is performed according to the determination result, and a bit string is output.

More specifically, the difference unit 103 includes a target block (encoding target block) output from the image memory unit 101 in units of macroblocks and a reference block output from a motion prediction unit 109 described later. The pixel value is subtracted, and the prediction error of the subtraction result is output to DCT section 104.
The DCT unit 104 calculates the prediction error from the difference unit 103 or
DCT is performed on the target block input from the image memory unit 101 without passing through the difference unit 103. As a result, the prediction error and the target block are converted into a DCT block composed of a set of DCT coefficients.

The quantizing section 105 quantizes each DCT coefficient of the DCT block from the DCT section 104, and
Skip quantizer 120 for a quantized block composed of T coefficients
Output to The skip determination unit 120 determines whether or not the macroblock is a skip macroblock based on the quantized block and the motion vector, and outputs a determination result to the motion vector replacement unit 121 and the variable length coding unit 110. The skip determination unit 120 according to the present embodiment is configured to determine more macroblocks as skip macroblocks as compared with the conventional encoding device. Although a detailed determination method will be described with reference to FIG. 2, the skip determination unit 120 determines whether or not to be a skipped macroblock based on whether or not the quantization coefficient and the motion vector satisfy predetermined conditions. Here, the predetermined condition is set looser than the predetermined condition in the conventional encoding device,
As a result, the number of macroblocks determined to be skipped macroblocks is increased as compared with the related art, the code amount is reduced, and the compression ratio is improved as compared with the related art.

The part consisting of the inverse quantization unit 106, the inverse DCT unit 107 and the addition unit 108 performs the operation opposite to the operation of the difference unit 103, the DCT unit 104, and the quantization unit 105 so that the quantization block becomes the target. Decrypt the block. The decoded target block is held in the internal memory of the motion prediction unit 109, and is used by the difference unit 103 to calculate a prediction error.

The motion prediction unit 109 stores the decoded target block in the internal memory in frame units, reads a reference block from the internal memory in accordance with the motion vector output from the motion vector detection unit 102, and sends the reference block to the difference unit 103. Output. The variable length coding unit 110 performs variable length coding on the quantized block, the motion vector, and various types of header information, and outputs a bit string to the buffer 111. The various header information is information that defines various encoding methods for macroblocks and pictures, and is defined by the control unit 113. When the skip determination unit 120 notifies the skip determination unit 120 that the macroblock to be coded is a skip macroblock, the variable length coding unit 110 performs variable length coding of the quantized block, the motion vector, and various header information of the macroblock. Is omitted.

The buffer 111 has a variable length encoding unit 110
The output bit string is accumulated and output to the outside via the output terminal 112. The motion vector replacement unit 121 replaces the value of the motion vector with the value for the skipped macroblock for the macroblock determined to be the skipped macroblock by the skip determination unit 120, and outputs the value to the variable length coding unit 110. Details will be described later with reference to FIG. <Detailed Description of Skip Determining Unit 120> FIG. 2 is a flowchart showing the processing of the skip determining unit 120 in detail.

The skip judging section 120 executes the processing shown in the figure for each macroblock, that is, every time a quantized block is input from the quantizing section 105. Skip determination unit 1
20, when a quantization block is input from the quantization unit 105 and a motion vector is input from the motion vector detection unit 102,
It is determined whether or not the macroblock is an I picture (step 20).

When it is determined that the macroblock is not an I-picture, that is, when the quantized block is a B-picture or a P-picture, the skip determining unit 120 determines that the macroblock is the first or last macroblock of a slice. It is determined whether or not (step 21). The slice is MP for realizing random access during decoding.
Specified by the EG standard, random access is not realized if the first or last macroblock is skipped and omitted. For this reason, the skip determination unit 120 previously excludes the top and last macroblocks of the slice from skip macroblock targets in step 21 so as not to determine them as skip macroblocks.

If the result of the determination indicates that the macroblock is not the first or last macroblock of the slice, skip determination section 120 determines whether the quantized block of the macroblock is 0, that is, the quantized block in the quantized block. It is determined whether the conversion coefficients are all 0 (step 22). As a result of the determination, when it is determined that the quantized block is 0, the skip determination unit 120 determines whether the macroblock is a P picture or a B picture (Step 23).

If it is determined that the picture is a P-picture, the skip determining unit 120 calculates the magnitude of the motion vector of the macroblock, and determines whether or not the value is equal to or smaller than the threshold a (step 24). ). Here, it is assumed that the motion vector includes an x component and a y component, and the skip determination unit 120 performs a threshold determination for each of the x component and the y component. That is, the skip determination unit 12
A value of 0 determines whether both the x component and the y component are less than or equal to the threshold value a, or whether one or both of the x component and the y component are greater than the threshold value a.

When the skip determining unit 120 determines that the magnitude of the motion vector is equal to or smaller than the threshold value a, that is, when both the x component and the y component of the motion vector
If it is determined that it is equal to or less than a, the macro block is set as a skip macro block (step 25). As described above, in the case of a P picture, the skip determination unit 120 determines that the quantization block is 0, not the head or last macroblock of the slice, and the size of the motion vector is x and y components. Is less than or equal to threshold a,
It is determined that the macro block is a skip macro block. Here, in this embodiment, the threshold value a is 0.
5 is assumed. That is, the skip determination unit 120 determines whether the magnitude of the motion vector of the macroblock is 0.5 or less for both the x component and the y component, and determines that the macroblock is a skip macroblock when the magnitude is 0.5 or less. If it is 0.5 or less, even if the encoding of the macroblock is omitted,
The difference between the original image and the reproduced image is hardly noticeable in human vision, so there is practically no problem.

On the other hand, if it is determined in step 23 that the picture is a B picture, the skip determination unit 120 calculates a difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock, and determines that the difference is a threshold b. It is determined whether or not the following is true (step 26). Here, the immediately preceding macroblock means a macroblock that was a coding target immediately before the coding target macroblock. On a frame, it is usually a macroblock located to the left of the encoding-target macroblock. The difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock is calculated for each of the x component and the y component.
Is to perform threshold determination for each of the x and y components of the difference. That is, the skip determination unit 120
Is that both the x component and the y component of the difference are equal to or smaller than the threshold value b, or one or both of the x component and the y component are
Determine if it is greater than. Note that the determination in step 26 is that the direction of prediction (forward prediction, backward prediction, bidirectional prediction, etc.) is the same between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock, that is, It shall include that the frames are the same.

If it is determined in step 26 that the difference is equal to or smaller than the threshold value b for both the x component and the y component, the skip determination unit 120 determines that the macro block is a skip macro block (step 25). The skip determining unit 120 notifies the variable-length coding unit 110, the motion vector replacing unit 121, and the adding unit 108 that the macroblock has been determined to be a skipped macroblock.

As described above, in the case of a B picture, the skip determining unit 120 determines whether the motion vector is not the macroblock at the head or end of the slice, the quantization block is 0, and the motion vector of the immediately preceding macroblock. The difference is x
If both the component and the y component are equal to or smaller than the threshold b, it is determined that the macroblock is a skipped macroblock. Here, the threshold value b is set to 1.0 in the present embodiment. That is, the skip determination unit 12
If the difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock is determined to be 0 or less, the macroblock is determined to be a skip macroblock. If it is 1.0 or less, as in the case of the P picture, even if the coding of the macroblock is omitted, the difference between the original image and the reproduced image is almost inconspicuous to human eyes, so that there is a substantial problem. There is no. The threshold a and the threshold b are 0.5 and 1.0.
However, the present invention is not limited to this. For example, the user may be able to set an arbitrary value in the skip determination unit 120 within a range in which deterioration of image quality is allowable. <Detailed Description of Motion Vector Replacement Unit 121> FIG. 3 is a flowchart showing the processing of the motion vector replacement unit 121 in detail.

The motion vector replacement unit 121 executes the processing of the flowchart shown in FIG. 3 for each macroblock, that is, each time a motion vector for each macroblock is input from the motion vector detection unit 102. When the motion vector replacement unit 121 receives a notification from the skip determination unit 120 that the macroblock to be coded is to be a skipped macroblock (step 30), the motion vector replacement unit 121 determines whether the macroblock is a P-picture (that is, a B-picture). Is determined (step 31).

On the other hand, if the motion vector replacement unit 121 does not receive a notification from the skip determination unit 120 in step 30 that the macroblock to be encoded is a skip macroblock, the motion vector detection unit 1
02 is output to the variable-length encoding unit 11
0 is output (step 34). If it is determined in step 31 that the macroblock is a P-picture, the motion vector replacement unit 121 replaces the motion vector output from the motion vector detection unit 102 with 0 having no direction or magnitude with the variable length coding unit. Output to 110 (step 32).

When it is determined in step 31 that the macroblock is a B picture, the motion vector replacement unit 121 replaces the motion vector output from the motion vector detection unit 102 with the one of the macroblock.
The motion vector of the previously encoded macroblock is output to the variable length encoding unit 110 (step 33). Here, the motion vector replacement unit 121 has an internal memory (not shown) and holds the motion vector of the macroblock immediately before the macroblock to be coded in the B picture. That is, in step 33, the motion vector replacement unit 121 replaces the motion vector of the macroblock output from the motion vector detection unit 102 with the motion vector held in the internal memory, that is, the one immediately before the macroblock. Is output as the motion vector of the macroblock encoded in.

The motion vector replacement unit 121 determines in step 3
In step 3, the motion vector output to the variable length coding unit 110 is stored in the internal memory. Also, a motion vector replacement unit 1
21 holds the motion vector output to the variable length coding unit 110 in step 34 in the internal memory for the case of a B picture (step 35). Here, the motion vector held in the internal memory is such that the macroblock to be encoded next to the macroblock is the skip determination unit 12.
It is used when the motion vector replacement unit 121 performs the process of step 33 because it is determined to be a skipped macroblock by 0.

Here, the motion vector replacement unit 12
1 stores the motion vector in the internal memory, but the motion vector detection unit 102 stores the motion vector, and the motion vector replacement unit 121 reads out and uses the motion vector from the motion vector detection unit 102 when performing the process of step 33. May be configured. As described above, when the macroblock to be encoded is a skipped macroblock, the motion vector replacement unit 121 replaces the motion vector of the macroblock with the motion vector for the skipped macroblock and outputs it to the variable length coding unit 110. .

The variable length coding unit 110 notifies the skip determination unit 120 that the macroblock to be coded is a skipped macroblock, and the motion vector replacement unit 12
When the replaced motion vector is received from 1, the variable-length coding of the quantized block of the macroblock to be coded and its various header information is omitted. As described above, the encoding device 10 of the present embodiment sets the number of skipped macroblocks to be larger than that of the conventional encoding device by setting the determination condition for determining a skipped macroblock to be looser than that of the conventional encoding device. Therefore, there is an effect that the code amount after the variable length coding is reduced and the compression ratio is increased. It should be noted that if the conditions for the determination are loosened in this way, there is a concern that the image quality will be degraded when the original image and the reproduced image are compared. However, in step 24 and step 26 in FIG.
5, if the threshold value b is 1.0, in the reproduced image,
Deterioration of the skipped macroblock portion is insignificant to human vision and poses no problem.

The coding apparatus 10 can realize higher quality coding by allocating the allocated bit amount reduced by the macroblock skip to the coding of another macroblock. <Second Embodiment> An encoding device 20 according to a second embodiment will be described below.
Will be described.

FIG. 4 is a block diagram showing the configuration of the encoding device 20 according to the second embodiment. The encoding device 20 of FIG.
The encoding device 10 is different from the encoding device 10 in that a feature extraction unit 222 is provided, and a skip determination unit 220 is provided instead of the skip determination unit 120. Other configurations are the same. Hereinafter, the feature extracting unit 222 and the skip determining unit 220 will be mainly described. <Feature Extraction Unit 222> The feature extraction unit 222 calculates, for each macroblock stored in the image memory unit 101, the variance (A) of each of the four blocks of the luminance signal in the macroblock.
CT), and the difference D between the maximum value (ACTmax) and the minimum value (ACTmin) is calculated by the skip determination unit 2
20.

Specifically, the feature extracting unit 222 calculates the variance by the following equation.

[0038]

(Equation 1) Here, Yi (i = 1 to 64) represents a value of 64 luminance signals in a block of 8 pixels × 8 lines, Yave represents an average value of 64 luminance signals, and ＾ 2 represents a square. That is, the variance is
It is represented by the sum of squares of the difference between each luminance signal value and the average value.

The feature extracting unit 222 calculates (Equation 1) for each of the four blocks to calculate four variances. Then, by comparing them, ACTmax and ACTmin are determined from the four, and the difference D is determined.
(= ACTmax−ACTmin) is calculated and output to the skip determination unit 120. Here, the variance has a large value when the luminance difference in the image is large, for example, when an edge or a boundary is included, and decreases when the luminance difference in the image is small. Therefore, it can be said that the macroblock image includes an image having a large difference in luminance and a small image when the value of the difference D is large, and it can be said that the luminance difference in each block is almost uniform when the value of the difference D is small. Pixel shift is more conspicuous in a macroblock including an image having a large luminance difference and an image having a small luminance difference than an image having a substantially uniform luminance difference. Therefore, the difference D
Is an index indicating the degree of conspicuousness of a pixel shift due to omission of a macroblock, and an index indicating a degree of deterioration of image quality. <Skip Determination Unit 220> The skip determination unit 220 determines whether the difference D output from the feature extraction unit 222 is large, whether the quantization coefficients are all 0, and whether the motion vector satisfies a predetermined condition. It is determined whether or not to perform a block skip.

FIG. 5 is a flowchart showing the details of the processing of the skip determining section 220. This figure differs from FIG. 2 in that the same step numbers as those in FIG. 2 indicate the same processing, and steps 54 and 56 are used instead of steps 24 and 26. Hereinafter, step 54
And step 56 will be described. In step 54, the skip determination unit 220 first calculates the magnitude of the motion vector of the macroblock,
The difference D is received from 22. Next, the skip determination unit 220
Determines whether the magnitude of the calculated motion vector is 0 for both the x component and the y component. As a result of this determination, when both the x component and the y component are 0, the skip determination unit 220 sets the macroblock as a skip macroblock.
On the other hand, as a result of the determination, the magnitude of the motion vector is x component, y
When both components are not 0, the skip determination unit 220 determines whether the magnitude of the motion vector is greater than 0 for both the x component and the y component and is equal to or less than the threshold value a and the difference D is equal to or less than the threshold value Ta. Is determined. As a result of this determination, when the magnitude of the motion vector is greater than 0 for both the x component and the y component and is equal to or less than the threshold value a and the difference D is equal to or less than the threshold value Ta, the skip determination unit 220 Let the block be a skip macroblock.

In step 56, the skip determining section 22
0 first calculates the difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock, and receives the difference D from the feature extraction unit 222. Next, the skip determination unit 220 determines whether the calculated difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock is 0 for both the x component and the y component. As a result of this determination, when both the x component and the y component are 0, the skip determination unit 220 sets the macroblock as a skip macroblock. On the other hand, if the result of the determination is that the difference is not 0 for both the x component and the y component, the skip determination unit 220
Determines whether the difference is greater than 0 for both the x component and the y component and equal to or less than the threshold value b, and whether the difference D from the feature extraction unit 222 is equal to or less than the threshold value Ta. As a result of this determination, if the difference is greater than 0 for both the x component and the y component and is equal to or less than the threshold value b and the difference D from the feature extraction unit 222 is equal to or less than the threshold value Ta, the skip determination unit 220 Makes the macroblock a skip macroblock.

Here, in steps 54 and 56, the threshold values a and b are the same as in the first embodiment. The threshold value Ta is 256. If the difference D is smaller than 256, the pixel shift of the image is almost inconspicuous when the motion vector of the skipped macroblock is replaced, and human eyes cannot recognize the deterioration of the image quality. In addition,
The threshold value Ta is not limited to 256. <Third Embodiment> An encoding device 30 according to a third embodiment will be described below.
Will be described.

The encoding apparatus 30 of the third embodiment differs from the encoding apparatus 20 shown in FIG.
2 and the feature extraction unit 322 instead of the skip determination unit 220
(Not shown) and skip determination unit 320 (not shown)
Is different from the encoding device 20. The feature extraction unit 322 calculates the variance (ACT) of each of the four blocks of the luminance signal in the macroblock by (Equation 1) and compares them to obtain the maximum value (ACTma).
x) and the minimum value (ACTmin) are determined, and ACTma
The difference D between x and ACTmin is calculated. Then, the feature extraction unit 322 outputs the ACTmax and the difference D to the skip determination unit 320.

The skip judging section 320 performs the following steps 64 and 66 instead of steps 54 and 56 in the flowchart shown in FIG. That is, in step 64, the skip determination unit 32
0 first calculates the magnitude of the motion vector of the macroblock, and receives ACTmax and the difference D from the feature extraction unit 322. Next, the skip determination unit 320 determines whether the magnitude of the calculated motion vector is 0 for both the x component and the y component. As a result of this determination, when both the x component and the y component are 0, the skip determination unit 220 sets the macroblock as a skip macroblock. On the other hand, as a result of the determination, when the magnitude of the motion vector is not 0 for both the x component and the y component, the skip determination unit 320 determines that the magnitude of the [1] motion vector is greater than 0 for both the x component and the y component. [2] Difference D is equal to threshold value Ta
[3] ACTmax is equal to threshold value Tb
It is determined whether or not: As a result of the determination, the above [1]
If both [2] and [3] are positively determined, the skip determination unit 320 sets the macroblock as a skipped macroblock.

In step 66, the skip determining section 32
0 calculates the difference between the motion vector of the macroblock and the motion vector of the immediately preceding macroblock, and receives the difference D and ACTmax from the feature extraction unit 322.
Next, the skip determination unit 320 determines whether the difference between the calculated motion vector of the macroblock and the motion vector of the immediately preceding macroblock is 0 for both the x component and the y component. As a result of this determination, when both the x component and the y component are 0, the skip determination unit 320 sets the macro block as a skip macro block. On the other hand, as a result of the determination, if the difference is not 0 for both the x component and the y component, the skip determination unit 320 [4] determines whether the difference is greater than 0 for both the x component and the y component and equal to or less than the threshold b. [5] Whether the difference D from the feature extraction unit 322 is equal to or smaller than the threshold value Ta,
[6] It is determined whether ACTmax is equal to or smaller than a threshold value Tb. Here, the determination conditions of [5] and [6] are the same as [2] and [3] in step 64.
As a result of this determination, when both [4], [5], and [6] are positively determined, the skip determination unit 320 sets the macroblock as a skipped macroblock.

In steps 64 and 66, the threshold values a and b are the same as in the first embodiment, and the threshold value Ta is 256, the threshold value is 256 as in the second embodiment, and the threshold value Tb is 1024. . The threshold values Ta and Tb are not limited to 256 and 1024. AC
A macroblock image in which Tmax is smaller than 1024 and the difference D is smaller than 256 indicates that the luminance change is smaller than the other macroblock images and that the luminance change in the image is smaller. Has such a characteristic that the luminance change is small in the entire frame and it is difficult to recognize a pixel shift of an image having a small luminance change in a macroblock. Using this characteristic,
The skip determination unit 320 determines that such a macroblock of an image is a skip macroblock, and that other macroblocks, that is, macroblocks of an image having a large luminance change within the entire frame or macroblock, are not skip macroblocks. .

With this configuration, the encoding device 30 has the effect of improving the compression efficiency because macroblocks of an image whose image quality is hardly discernible to human eyes are skipped macroblocks. <Fourth Embodiment> An encoding device 40 according to a fourth embodiment will be described below.
Will be described.

The encoding apparatus 40 of the fourth embodiment is different from the configuration of the encoding apparatus 20 shown in FIG.
And a feature extracting unit 422 instead of the skip determining unit 220
(Not shown) and skip determination unit 420 (not shown)
This is different from the encoding device 20 in that The feature extraction unit 422 calculates ACTmax and ACT according to (Equation 1).
and the difference D is calculated. Also, the feature extraction unit 422 calculates the average value YM of the luminance for each macroblock.
Bave, that is, an arithmetic average is calculated by dividing the sum of the luminance values in the macroblock by the number of luminances. Feature extraction unit 42
2 outputs the difference D and the average value YMave to the skip determination unit 420.

The skip determination unit 420 performs the following steps 74 and 76 immediately before the processing of steps 54 and 56 in the flowchart shown in FIG. The other processing is as shown in FIG. 5, and therefore the following description will focus on step 54 and step 56. That is, in step 74, skip determination section 420 outputs YMBav output from feature extraction section 422.
It is determined whether or not e is equal to or less than the threshold value c. Here, the threshold value Tc is 128. The threshold value Tc is not limited to 128. As a result of the determination, YMave
Is equal to or smaller than the threshold value Tc, the skip determination unit 420
Multiplies the threshold value a for determining the magnitude of the motion vector used in the processing of step 54 following step 74 by n times (where n> 1). On the other hand, skip determination section 420
If the average value YMave is larger than the threshold value Tc, the process proceeds to step 54 without performing n times the threshold value a. Here, n is 4, for example.

In step 54, when it is determined in step 74 that YMBave is equal to or smaller than threshold value Tc, skip determination unit 420 determines that threshold value a × n is n times as large as threshold value a. (That is, 2.0 when the threshold a is 0.5 and n is 4), and when it is determined that YMBave is larger than the threshold Tc, the threshold a is The judgment is made using the above.

In step 76, the skip determining section 42
A value of 0 determines whether the YMave output from the feature extraction unit 422 is equal to or smaller than the threshold value Tc. If the result of the determination is that YMBave is equal to or smaller than the threshold value Tc, the skip determination unit 420 multiplies the threshold value b for determining the magnitude of the difference between the motion vectors used in the processing of step 56 following step 76 by n. On the other hand, the skip determination unit 42
If the average value YMave is larger than the threshold value Tc, the process proceeds to step 56 without performing n times the threshold value b.

Accordingly, in step 56, when it is determined in step 76 that YMave is equal to or smaller than threshold value Tc, skip determination unit 420 determines that threshold value b × n, which is n times threshold value b. (That is, 4.0 when threshold b is 1.0 and n is 4), and when it is determined that YMBave is larger than threshold Tc, threshold b The determination is made using

Thus, skip determining section 420 determines that more macroblocks are skipped macroblocks in a macroblock having a smaller average luminance, that is, a darker image. Macroblocks of a dark image are less noticeable to human visual perception than macroblocks of a bright image. Therefore, when the image of the macroblock is dark, the encoding device 40 of the present embodiment increases the threshold value a or the threshold value b used for determining the threshold value of the motion vector to determine the dark image. In this case, more skipped macroblocks are generated in the macroblock, thereby increasing the compression efficiency.

Although the encoding apparatus according to the present invention has been described based on the embodiments, the present invention is not limited to these embodiments, and may be as follows. (1) Encoding devices 10 to 40 in first to fourth embodiments
Performs inter-frame coding, but may use inter-field coding instead of inter-frame coding, or may be configured to use both inter-frame coding and inter-field coding. . (2) In the fourth embodiment, the skip determination unit 420
Although steps 74 and 76 are performed immediately before the processing of steps 54 and 56 in the second embodiment, steps 74 and 76 are performed immediately before the processing of steps 64 and 66 in the third embodiment. It may be. (3) In the first to fourth embodiments, the function of each component is programmed and recorded in a ROM, and a CPU, a RAM, a ROM
May be realized by a microcomputer comprising:
More specifically, the CPU reads out the program from the ROM and executes the program to first execute a step of calculating an index indicating the degree of conspicuousness of a pixel shift when encoding of the encoding target macroblock is omitted. Next, the CPU executes a step of omitting the coding of the macroblock to be coded according to the index, that is, determining the macroblock as a skipped macroblock and omitting it.

More specifically, for example, as a program corresponding to the first embodiment, the CPU calculates, as indices, the magnitude of the motion vector of the macroblock and the difference from the motion vector of the immediately preceding macroblock. In a picture, when the quantized DCT coefficient of the macroblock is 0 and the magnitude of the motion vector is less than or equal to the threshold a in both the x component and the y component, in the B picture, the quantized When the DCT coefficient is 0 and the difference from the immediately preceding motion vector is less than or equal to the threshold b for both the x component and the y component, a program that is determined to be a skipped macroblock and omitted is read from the ROM and executed. (4) The program described in (3) above can be recorded on a recording medium or distributed and distributed via various communication paths. Such a recording medium includes an IC card, an optical disk, a flexible disk, a ROM, and the like. A computer program for causing a general-purpose computer or a household electric appliance having a program execution function to execute a recording operation procedure (the procedure shown in FIG. 6) of the digital data recording control system described in the present embodiment is recorded on a recording medium. , Or can be distributed and distributed via various communication channels. Such a recording medium includes an IC card, an optical disk, a flexible disk, a ROM, and the like. The distributed and distributed computer program is provided for use by being installed on a home appliance or a computer having a program execution function, and the home appliance and the computer execute the computer program.
A digital data recording control system as described in the present embodiment is realized. (5) In the fourth embodiment, the skip determination unit 420
When YMBave is equal to or smaller than the threshold value Tc, the threshold value a and the threshold value b are multiplied by n instead of the threshold value a and the threshold value b at the time of the threshold value determination in the steps 54 and 56. The determination is made using the threshold values a × n and the threshold value b × n obtained as described above. However, the determination is not limited to n times, and for example, the same value may be uniformly used for the threshold value a and the threshold value b. The determination may be performed using the added value.

[0056]

The encoding apparatus according to the present invention is an encoding apparatus for performing encoding including motion compensation prediction on a current block for each block of moving image data. And a control unit that controls to omit encoding of the encoding target block according to the index.

According to this configuration, the encoding device encodes a block in which the pixel shift is conspicuous, and omits the coding of the block in which the pixel shift is not conspicuous. Even if a block has a small displacement, the block is encoded without being omitted, which eliminates the mismatch between the degree of pixel displacement and the omission of encoding, and has the effect of reducing the amount of encoding processing and increasing the compression ratio. .

In addition, the calculating means includes:
The magnitude of the motion vector obtained by motion compensation prediction and the difference between the motion vector of the current block and the block coded immediately before the current block are calculated. If the magnitude is larger than 0 and smaller than the first threshold, the encoding of the current block is omitted in bidirectional prediction encoding if the difference is larger than 0 and smaller than the second threshold. With this control, the encoding result of the encoding target block is not output.

The pixel shift amount when the coding is omitted increases as the magnitude of the motion vector and the difference between the motion vectors increase. Therefore, in the coding apparatus having the above configuration, the control unit makes the determination using the first threshold value and the second threshold value so that the pixel shift amount when the coding is omitted is smaller than the predetermined shift amount. Is omitted, and the block in which the pixel shift is larger than a predetermined shift amount is coded without omission.
As described above, in the present encoding apparatus, the block having a small pixel shift coincides with the block from which encoding is omitted, and thus has an effect of improving the encoding efficiency.

The calculating means calculates a magnitude of the motion vector obtained by the motion compensation prediction, and a first calculating unit for calculating a difference between the motion vector of the block to be coded and the block coded immediately before the current block. A second calculating unit that calculates a variance of luminance in a small block obtained by dividing the current block into a plurality of blocks, and calculates a difference between a maximum value and a minimum value of the variance, wherein the index is a magnitude of the motion vector. And the difference between the maximum value and the minimum value of the variance, and the control unit determines that the magnitude of the motion vector is larger than 0 in unidirectional prediction encoding. If the difference is larger than 0 and smaller than a second threshold value in bidirectional prediction encoding, the first block is determined to be a candidate for coding omission when the difference is larger than 0 and smaller than a second threshold value. And a second determining unit that determines the current block to be a candidate for coding omission when the difference calculated by the second calculating unit is smaller than a third threshold value. When both of the two determination units determine that the candidate for encoding is to be omitted, control is performed such that encoding of the current block is omitted.

According to this configuration, the encoding apparatus omits the encoding of a block where the difference in variance within the block is small, that is, the luminance difference is uniform and the pixel deviation due to the omission of encoding is less noticeable. Since the coding is omitted according to the conspicuousness, that is, the degree of deterioration, the coding processing efficiency and the compression efficiency are improved. Further, the second determination unit is configured such that, when the difference calculated by the second calculation unit is smaller than a third threshold value, and when the maximum value of the variance calculated by the second calculation unit is smaller than a fourth threshold value. When the size is smaller, the current block to be coded is determined to be a candidate for coding omission.

According to this configuration, the encoding apparatus can generate an image in which the difference between the maximum value and the minimum value of the luminance variance is smaller than the third threshold value and the maximum value of the luminance variance is smaller than the fourth threshold value; That is, encoding of a flatter image in which the luminance change in the block is small is omitted. As described above, in the case of a flatter image having a small change in luminance, the pixel shift due to the omission of encoding is not conspicuous. Therefore, the encoding is omitted according to the conspicuousness of the pixel shift, and the encoding processing efficiency and the compression efficiency are reduced. Is effective.

The calculating means further includes a third calculating section for calculating an average value of the luminance of the current block as a part of the index, and the control means further includes a third determining section. Before the determination, it is determined whether or not the average value is smaller than a fifth threshold value. If it is determined that the average value is smaller than the first threshold value and the second threshold value, each carry is performed. A threshold varying unit that causes the first determination unit to make a determination using the obtained value.

According to this configuration, the encoding apparatus carries out the determination by increasing the first threshold value and the second threshold value in the case of a block of an image darker than a certain threshold value, and thereby performs the determination of the dark image. The probability that the coding of the block is omitted is increased. Since dark images have the characteristic that image quality deterioration due to omission of encoding is difficult to be recognized by human eyes, this configuration is used to generate more blocks whose encoding is omitted in darker image blocks. However, since a block having a low degree of image deterioration corresponds to a block for which encoding is omitted, there is an effect that encoding processing efficiency and compression efficiency are good.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an encoding device 10.

FIG. 2 is a flowchart illustrating a process of a skip determination unit 120 in detail.

FIG. 3 is a flowchart illustrating a process of a motion vector replacement unit 121 in detail.

FIG. 4 is a block diagram illustrating a configuration of an encoding device 20 according to a second embodiment.

FIG. 5 is a flowchart showing details of a process of a skip determination unit 220;

[Explanation of symbols]

 Reference Signs List 10 Encoding device 100 Input terminal 101 Image memory unit 102 Motion vector detection unit 102 103 Difference unit 104 DCT unit 105 Quantization unit 106 Inverse quantization unit 107 Inverse DCT unit 108 Addition unit 109 Motion prediction unit 110 Variable length encoding unit 111 Buffer 112 Output terminal 113 Control unit 120 Skip determination unit 121 Motion vector replacement unit 220 Skip determination unit 222 Feature extraction unit

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5C059 KK00 KK02 LC09 MA00 MA05 MA23 ME01 NN10 NN11 NN21 NN28 NN36 PP04 PP06 PP07 SS20 TA16 TB08 TC01 TC10 TC11 TC12 TC24 TD02 TD04 TD05 TD12 5J064 AA02 BA01 BA09 BA03 BC01 BC01

Claims (7)

[Claims] An encoding apparatus for performing encoding including motion compensation prediction on an encoding target block for each block of moving image data, wherein a pixel shift occurs when encoding of the encoding target block is omitted. An encoding apparatus comprising: a calculating unit that calculates an index indicating the degree of conspicuousness of the target block; and a control unit that controls to omit encoding of the current block according to the index. 2. The calculation means calculates, as the index, a magnitude of a motion vector obtained by motion compensation prediction, and a difference between a motion vector of an encoding target block and a block encoded immediately before the current block. The control means, when the magnitude is larger than 0 and smaller than the first threshold value in unidirectional prediction encoding, the difference is larger than 0 in bidirectional prediction encoding.
2. The encoding apparatus according to claim 1, wherein when the value is smaller than the threshold value, the encoding result of the encoding target block is not output by controlling to omit encoding of the encoding target block. 3. A first calculating unit for calculating a magnitude of a motion vector obtained by motion compensation prediction and a difference between a motion vector of an encoding target block and a block encoded immediately before the encoding target block. And a second calculating unit that calculates a variance of luminance in a small block obtained by dividing the current block into a plurality of blocks, and calculates a difference between a maximum value and a minimum value of the variance. A magnitude, a difference between the motion vectors, and a difference between a maximum value and a minimum value of the variance, wherein the control unit determines that the magnitude of the motion vector is greater than 0 and larger than 0 in the unidirectional prediction encoding. When the difference is larger than 0 and smaller than a second threshold value in bidirectional prediction encoding when the difference is smaller than the threshold value, the first determination is made to determine that the current block is a candidate for encoding omission. And a second determining unit that determines the current block to be a candidate for coding omission when the difference calculated by the second calculating unit is smaller than a third threshold value. 2. The control unit according to claim 1, wherein when both of the two determination units determine that the candidate for encoding is to be omitted, encoding of the current block is omitted.
An encoding device according to claim 1. 4. The method according to claim 1, wherein the difference calculated by the second calculation unit is smaller than a third threshold value, and the maximum value of the variance calculated by the second calculation unit is set to a fourth value. The encoding apparatus according to claim 3, wherein the encoding target block is determined as a candidate for encoding when the encoding target block is smaller than the threshold value. 5. The calculation unit further includes: a third calculation unit that calculates, as a part of the index, an average value of luminance of an encoding target block, the control unit further includes: the first determination unit. It is determined whether or not the average value is smaller than a fifth threshold value before the determination by (1), and if it is determined that the average value is smaller, each of the first and second threshold values is replaced by a digit. The encoding device according to claim 3 or 4, further comprising a threshold variable unit that causes the first determination unit to make a determination using the increased value. 6. An encoding method for performing encoding including motion compensation prediction on an encoding target block for each block of moving image data, wherein a pixel shift occurs when encoding of the encoding target block is omitted. An encoding method comprising: a calculating step of calculating an index indicating the degree of conspicuousness of a target block; and a control step of performing control so as to omit encoding of a current block according to the index. 7. A computer-readable recording medium in which a program for causing a computer to execute coding including motion compensation prediction on a block to be coded for each block of moving image data is recorded. Calculates a step of calculating an index indicating the degree of conspicuousness of pixel shift when encoding of the encoding target block is omitted, and controls the computer to omit encoding of the encoding target block according to the index. A recording medium having recorded thereon a program for executing a control step.